High operating frequencies are desirable for power conditioning equipment (such as power supplies, power amplifies and lamp ballasts) in order to reduce the size of magnetic components. Unfortunately, however, it has been heretofore impossible to substantially reduce the size of magnetic components because the operating flux density has to be reduced too much to enable proper operation and still have acceptable losses at high frequencies. In particular, while at low frequencies (e.g. 100 kHz), the operating flux density can be in the 1000-3000 Gauss range using semi-conducting ferromagnetic materials (typically ferrites), the operating flux density to achieve acceptable losses in the megahertz range is from 20 to 300 Gauss, which is too low to give any significant size advantage.
In addition, there is considerable drive to push the operating frequencies of power conditioning equipment to frequencies even higher than 5.0 MHz. However, the losses of the current high frequency materials rise very sharply with increasing frequency in 1 to 10 MHz range, making these material unsuitable for use at such high frequencies. In particular, the losses of these high frequency materials increase as the 4th power of the frequency above one megahertz.
A variety of publications describe the state of the art in this field. These publications include a book by A. Goldman: Modern Ferrite Technology (Van Nostrand Reinhold, NY, 1990). Detailed properties of soft ferrites are also described in a book by E. C. Snelling: Soft Ferries: Properties and Applications, 2nd edition (Butterworths, London, 1988).
The following are the four major reasons for the two problems described hereinabove and for making the current state-of-the-art materials unsuitable for use in high frequency and high flux density operation.
First, the people involved in the development of high frequency magnetic materials for use in power electronics have only tried to reduce eddy current losses. Unfortunately, the methods used to control eddy current losses (these methods include reducing the grain size to increase the grain boundary surface and increased addition of materials like SiO2 and CaO) have resulted in increased hysteresis loss. This increased hysteresis loss produces a very high dependence of the net high frequency losses on the flux density because hysteresis loss varies as the cube of the flux density. Thus, making these materials unsuitable for use at high flux densities and high frequencies.
Yet, another problem with the current approach for making low loss high frequency magnetic material is that the eddy current losses are not completely eliminated and they still make substantial contribution to the total loss in the 100 kHz to 1.0 MHz. This is evident from the measured net losses, which increase as the square of the frequency (hysteresis loss increase, generally, linearly with frequency). In addition, the eddy loss dominate the net losses at frequencies higher than 1 MHz. The main reason for this failure to control eddy current losses are very thin, non-uniform grain boundary layer, which is supposed to produce high resistivity. However, at high frequencies this layer begins to yield due to capacitive electrical shorting of boundary layer. The thinner that layer the more noticeable is this effect. The measured resistivity of the current state-of-the-art materials show a continuous decrease in the value of the resistivity in 100 kHz to 10 MHz range. All the previous attempts to reduce this effect have failed.
Yet, another problem with the current high frequency material is that very large electric fields are produced at the grain boundaries, which leads to a dielectric break down of the material at the grain boundaries and to a resistivity, which is a decreasing function of the flux density. This is because the thickness of the grain boundary layer is extremely small (for typical high frequency ferrite the thickness is in 10−3 to 10−2 micron range) and the fact that the resistivity of the crystallites is negligible compared to that of grain boundary layer. In particular, even for a very small core, the induced electric field at 1.0 MHz and at 500 Gauss exceeds 10 kV/cm. These values of electric field far exceed the dielectric strength (typically around 2 kV/cm) of the material at the grain boundary. This leads to resistivity, which decreases rapidly with increasing frequency and flux density. The end result is the increased dependence of high frequency loss on flux density, which further limits the flux density at which these materials can be operated.
The fourth major problem relate to the use of these materials for very high frequency applications and for very high power. This is because of large permeability and dielectric constant, which give a small electromagnetic wavelength at high frequencies. Due to this very small wavelength, dimensional resonance can be set up even in small cores and leads to additional losses. In particular, for typical high frequency ferrites, the wavelength is about 1 mm at 10 MHz and thus making these materials unsuitable even for very small cores. As power level of application such as power supply increases, the size of core increases making the problem worst and restricting the use of these materials to very low powers at such high frequencies.
Accordingly, it is desirable to provide high power magnetic material, which can be operated at high flux levels (e.g. greater than 500 Gauss) while maintaining high frequency operation (e.g. 1 to 10 MHz). Furthermore, such material should have a resistivity, which is independent of the frequency and flux density and should allow for independent control and reduction of hysteresis loss and eddy current loss. In addition, such magnetic material should have a structure, which allows for effective size of the core to be substantially smaller than the electromagnetic wavelength at high frequencies.